Influence of Family Demographic Factors on Social Communication Questionnaire (SCQ) Scores

Background: Screeners for autism spectrum disorders (ASDs) are important for determining ASD risk in population-based studies. Consequently, identifying socio-economic status (SES) factors that affect the validity of an ASD screener can improve the accurate identification of autism-related behaviors and inform future screening efforts. The Social Communication Questionnaire (SCQ) is a parent report screener for ASD that has been widely used in epidemiological studies.  It is unclear whether SES factors affect the performance of the SCQ, as studies have reported conflicting findings (Corsello et al, 2007; Tsai et al, 2012).  

Objectives: To examine the impact of maternal race, ethnicity, language, education and family income on SCQ scores.

Methods: SCQ data were collected in the Study to Explore Early Development-Phase I (SEED I), a multi-site (California, Colorado, Georgia, Maryland, North Carolina and Pennsylvania) case-control study exploring the phenotypes and determinants of ASD. Participants aged 2-5 years were recruited through organizations providing ASD-related services and from randomly selected vital records. Participants were screened using the SCQ to determine ASD risk. Children were classified as ASD, non-ASD developmental disability (DD), or population-comparison (POP) after screening and a clinical assessment for those with 11 or more points on the SCQ. Demographic data were collected in a standardized maternal interview. Separate linear regressions were run for children with final classification of ASD (n=667, SCQ mean = 17.2 ± 6.1), DD (n=990, SCQ mean = 7.2 ± 5.1) and POP (n=936, SCQ mean =4.3 ± 3.2).  In all models, family income (<10K, 10-30K, 30-50K, 50-70K, 70-90K, 90-110K, >110K), maternal education (less than high school, high school, some college/trade, bachelor degree or advanced degree), maternal race (White, Black, Asian, Native American/Native Alaskan/Native Hawaiian/Pacific Islander, Multi-racial), ethnicity (Hispanic, Non-Hispanic) and language in the home (Spanish or English) were regressed on total SCQ score (range 0-35).

Results: This work reports preliminary results for our study sample.  Higher SCQ scores were predicted in the ASD group for lower family income (β=-.60, p<0.001), in the DD group for lower family income (β =-.70, p<0.001) and less maternal education (β =-.90, p<0.001), and in the POP group for lower family income (β =-.32, p<0.001) and less maternal education (β =-.62, p<0.001). Maternal race and ethnicity and language spoken in the home did not have a significant influence on SCQ scores in any study group. Collinearity was not a concern in any of the models.

Conclusions: Our results indicate that SCQ scores are influenced by family demographics, particularly family income and maternal education.  Lower family income predicted higher SCQ scores for children in all three SEED study groups.  Moreover, less maternal education predicted higher SCQ scores in the DD and POP comparison groups.  These results indicate that family income and maternal education appear to be important considerations when interpreting SCQ scores; future revisions of the SCQ should take into account the influence of these demographic variables.  Future analyses could consider the sensitivity, specificity, positive predictive value, and negative predictive value of the SCQ when stratified by demographic variables.